fix: null value handling on PromQL's join (#6289)

* fix: null value handling on PromQL's join

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

* update sqlness comment

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>

---------

Signed-off-by: Ruihang Xia <waynestxia@gmail.com>
This commit is contained in:
Ruihang Xia
2025-06-10 21:24:09 +08:00
committed by GitHub
parent c26138963e
commit 2fabe346a1
5 changed files with 43 additions and 38 deletions

48
Cargo.lock generated
View File

@@ -3252,7 +3252,7 @@ dependencies = [
[[package]]
name = "datafusion"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"arrow 54.2.1",
"arrow-array 54.2.1",
@@ -3303,7 +3303,7 @@ dependencies = [
[[package]]
name = "datafusion-catalog"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"arrow 54.2.1",
"async-trait",
@@ -3323,7 +3323,7 @@ dependencies = [
[[package]]
name = "datafusion-catalog-listing"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"arrow 54.2.1",
"arrow-schema 54.3.1",
@@ -3346,7 +3346,7 @@ dependencies = [
[[package]]
name = "datafusion-common"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"ahash 0.8.11",
"arrow 54.2.1",
@@ -3371,7 +3371,7 @@ dependencies = [
[[package]]
name = "datafusion-common-runtime"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"log",
"tokio",
@@ -3380,12 +3380,12 @@ dependencies = [
[[package]]
name = "datafusion-doc"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
[[package]]
name = "datafusion-execution"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"arrow 54.2.1",
"dashmap",
@@ -3403,7 +3403,7 @@ dependencies = [
[[package]]
name = "datafusion-expr"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"arrow 54.2.1",
"chrono",
@@ -3423,7 +3423,7 @@ dependencies = [
[[package]]
name = "datafusion-expr-common"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"arrow 54.2.1",
"datafusion-common",
@@ -3434,7 +3434,7 @@ dependencies = [
[[package]]
name = "datafusion-functions"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"arrow 54.2.1",
"arrow-buffer 54.3.1",
@@ -3463,7 +3463,7 @@ dependencies = [
[[package]]
name = "datafusion-functions-aggregate"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"ahash 0.8.11",
"arrow 54.2.1",
@@ -3484,7 +3484,7 @@ dependencies = [
[[package]]
name = "datafusion-functions-aggregate-common"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"ahash 0.8.11",
"arrow 54.2.1",
@@ -3496,7 +3496,7 @@ dependencies = [
[[package]]
name = "datafusion-functions-nested"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"arrow 54.2.1",
"arrow-array 54.2.1",
@@ -3518,7 +3518,7 @@ dependencies = [
[[package]]
name = "datafusion-functions-table"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"arrow 54.2.1",
"async-trait",
@@ -3533,7 +3533,7 @@ dependencies = [
[[package]]
name = "datafusion-functions-window"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"datafusion-common",
"datafusion-doc",
@@ -3549,7 +3549,7 @@ dependencies = [
[[package]]
name = "datafusion-functions-window-common"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"datafusion-common",
"datafusion-physical-expr-common",
@@ -3558,7 +3558,7 @@ dependencies = [
[[package]]
name = "datafusion-macros"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"datafusion-expr",
"quote",
@@ -3568,7 +3568,7 @@ dependencies = [
[[package]]
name = "datafusion-optimizer"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"arrow 54.2.1",
"chrono",
@@ -3586,7 +3586,7 @@ dependencies = [
[[package]]
name = "datafusion-physical-expr"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"ahash 0.8.11",
"arrow 54.2.1",
@@ -3609,7 +3609,7 @@ dependencies = [
[[package]]
name = "datafusion-physical-expr-common"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"ahash 0.8.11",
"arrow 54.2.1",
@@ -3622,7 +3622,7 @@ dependencies = [
[[package]]
name = "datafusion-physical-optimizer"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"arrow 54.2.1",
"arrow-schema 54.3.1",
@@ -3643,7 +3643,7 @@ dependencies = [
[[package]]
name = "datafusion-physical-plan"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"ahash 0.8.11",
"arrow 54.2.1",
@@ -3673,7 +3673,7 @@ dependencies = [
[[package]]
name = "datafusion-sql"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"arrow 54.2.1",
"arrow-array 54.2.1",
@@ -3691,7 +3691,7 @@ dependencies = [
[[package]]
name = "datafusion-substrait"
version = "45.0.0"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=e104c7cf62b11dd5fe41461b82514978234326b4#e104c7cf62b11dd5fe41461b82514978234326b4"
source = "git+https://github.com/waynexia/arrow-datafusion.git?rev=12c0381babd52c681043957e9d6ee083a03f7646#12c0381babd52c681043957e9d6ee083a03f7646"
dependencies = [
"async-recursion",
"async-trait",

View File

@@ -116,15 +116,15 @@ clap = { version = "4.4", features = ["derive"] }
config = "0.13.0"
crossbeam-utils = "0.8"
dashmap = "6.1"
datafusion = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "e104c7cf62b11dd5fe41461b82514978234326b4" }
datafusion-common = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "e104c7cf62b11dd5fe41461b82514978234326b4" }
datafusion-expr = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "e104c7cf62b11dd5fe41461b82514978234326b4" }
datafusion-functions = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "e104c7cf62b11dd5fe41461b82514978234326b4" }
datafusion-optimizer = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "e104c7cf62b11dd5fe41461b82514978234326b4" }
datafusion-physical-expr = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "e104c7cf62b11dd5fe41461b82514978234326b4" }
datafusion-physical-plan = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "e104c7cf62b11dd5fe41461b82514978234326b4" }
datafusion-sql = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "e104c7cf62b11dd5fe41461b82514978234326b4" }
datafusion-substrait = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "e104c7cf62b11dd5fe41461b82514978234326b4" }
datafusion = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "12c0381babd52c681043957e9d6ee083a03f7646" }
datafusion-common = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "12c0381babd52c681043957e9d6ee083a03f7646" }
datafusion-expr = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "12c0381babd52c681043957e9d6ee083a03f7646" }
datafusion-functions = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "12c0381babd52c681043957e9d6ee083a03f7646" }
datafusion-optimizer = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "12c0381babd52c681043957e9d6ee083a03f7646" }
datafusion-physical-expr = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "12c0381babd52c681043957e9d6ee083a03f7646" }
datafusion-physical-plan = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "12c0381babd52c681043957e9d6ee083a03f7646" }
datafusion-sql = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "12c0381babd52c681043957e9d6ee083a03f7646" }
datafusion-substrait = { git = "https://github.com/waynexia/arrow-datafusion.git", rev = "12c0381babd52c681043957e9d6ee083a03f7646" }
deadpool = "0.12"
deadpool-postgres = "0.14"
derive_builder = "0.20"

View File

@@ -2444,7 +2444,7 @@ impl PromPlanner {
LogicalPlanBuilder::from(left)
.alias(left_table_ref)
.context(DataFusionPlanningSnafu)?
.join(
.join_detailed(
right,
JoinType::Inner,
(
@@ -2458,6 +2458,7 @@ impl PromPlanner {
.collect::<Vec<_>>(),
),
None,
true,
)
.context(DataFusionPlanningSnafu)?
.build()

View File

@@ -675,11 +675,16 @@ insert into cache_miss_with_null_label values
Affected Rows: 4
-- SQLNESS SORT_RESULT 3 1
-- null!=null, so it will returns the empty set.
tql eval (3, 4, '1s') cache_hit_with_null_label / (cache_miss_with_null_label + cache_hit_with_null_label);
++
++
+-------+------------+---------------------+---------------------------------------------------------------------------------------------------------------+
| job | null_label | ts | lhs.greptime_value / rhs.cache_miss_with_null_label.greptime_value + cache_hit_with_null_label.greptime_value |
+-------+------------+---------------------+---------------------------------------------------------------------------------------------------------------+
| read | | 1970-01-01T00:00:03 | 0.5 |
| read | | 1970-01-01T00:00:04 | 0.75 |
| write | | 1970-01-01T00:00:03 | 0.5 |
| write | | 1970-01-01T00:00:04 | 0.6666666666666666 |
+-------+------------+---------------------+---------------------------------------------------------------------------------------------------------------+
-- SQLNESS SORT_RESULT 3 1
tql eval (3, 4, '1s') cache_hit_with_null_label / ignoring(null_label) (cache_miss_with_null_label + ignoring(null_label) cache_hit_with_null_label);

View File

@@ -325,7 +325,6 @@ insert into cache_miss_with_null_label values
(4000, "write", null, 2.0);
-- SQLNESS SORT_RESULT 3 1
-- null!=null, so it will returns the empty set.
tql eval (3, 4, '1s') cache_hit_with_null_label / (cache_miss_with_null_label + cache_hit_with_null_label);
-- SQLNESS SORT_RESULT 3 1